How the EU AI Act will impact the Channel

The EU AI Act brought in earlier this year will present both challenges and opportunities for companies.

The European Union (EU) AI Act is the first legal framework for AI to be introduced by a major regulator.

Passed on 13th March 2024, the new legislation addresses the risks of using AI as well as ensuring the technology meets fundamental rights, safety and ethical principles within Europe and outside.

Given its relative infancy, the AI Act's impact on the Channel hasn't been fully realised yet, but the implications for resellers, MSPs and VARs could well be far-reaching.

Jeff Green, CEO of Elisha Telecom, said that the key areas where the Act will have an effect are whether emotion recognition in the workplace, behavioural manipulation and biometric recognition are deemed either high-risk or unacceptable use cases.

"While this may seem drastic, all businesses must inform callers that calls are being recorded for training purposes," said Green. "Therefore, if emotion/sentiment recognition is being used, this catch-all should still cover the customer and supplier for the time being, but I’m no legal expert.”

From the end user’s perspective, Green said that this level of transparency in AI-driven systems may be off-putting and a challenge for the Channel. But, he added that, just like when interactive voice responses were introduced, it will soon become the norm as customers get used to more automated interactions.

"Generative AI and chatbot transparency will be required," said Green. "Whether delivered via voice or text, consumers will have to be informed that they are interacting with AI. Those who supply AI solutions will also need to shoulder the responsibility of overseeing and managing individual solutions for businesses, training staff and customers and recording all interactions accurately.

"For example, channel resellers will be the ones responsible for ensuring generative AI deployments in chatbots are well tested and won’t throw out errant replies. This takes a slight veer from today’s deployments, where many resellers rely on their vendor’s support and in-platform technology to implement and manage systems, and I think it would be time well-spent if resellers already delivering AI agree on their internal best practices and establish a code of conduct for their services – a necessity of the Act."

New opportunities

Green said that despite the challenges, the AI Act will open up new opportunities for differentiation and revenue maximisation. He said that resellers can provide AI management and support, with increased focus on the pre-consultancy and ongoing management stages of chatbots quickly become value-adds for companies looking to diversify income streams and find a new niche.

"In summary, many of the questions that the Act raise ultimately come back to 'Where is the line?'" said Green. "This will probably take a few years to establish, but in the meantime, transparency is essential, both for end customers and internal operations. Establish your best practices until the line is drawn and don't worry too much about the risk or penalties. Anyone found to have their toes on the wrong side of the line will likely get a well-intentioned tap on the wrist rather than be made an example of."

Abdul Terry, manager of channel sales at Jamf, said, "Over the next couple of years, the EU's AI Act will set the template of how companies, even those based outside the EU, move forward and implement AI, similar to what was seen with GDPR. It will be the foundation for organisations looking to address challenges presented by AI adoption – for example, the risk of AI replacing the workforce."

Jamie Marshall, co-founder of Everon, added, "It's going to take some time to understand AI enough to apply the correct measures and policies to govern it. You only have to listen to the wildly different views of some of the tech leaders right now to see that we don't seem to have a single clear understanding of risks. This is going to be a global issue we will have to overcome and ideally need to have lots of input from the tech companies at the forefront of AI to ensure we don’t define policies that are outdated or not fit for purpose."